AWS : Solutions Architect Professional Exam — Part 4

A Quick Review for Cost Control

Pisit J.
Sum up As A Service
4 min readSep 12, 2021

--

Part 4: Cost Control (12.5% of exam)

  • Cost Optimization & Monitoring

1 — They have three different accounts for their environments: DEV, UAT and PROD, which are all part of the consolidated billing account. The PROD account has purchased 10 r4.16xlarge Reserved EC2 Instances in the us-west-2a.

Currently, there is no running EC2 instance in the PROD account because the application is not live yet but in the DEV account, there are 5 r4.16xlarge EC2 instances running in the us-west-2a. In the UAT account, there are also 5 r4.16xlarge EC2 instances running in the us-west-1a.

In this scenario, which account benefits the most out of the Reserved Instance pricing ?

Only the DEV account benefits from the Reserved Pricing.

Note — The consolidated billing feature of AWS Organizations, for billing purpose, treats all the accounts in the organization as one account. This means that all accounts in the organization can receive the hourly cost benefit of Reserved Instances that are purchased by any other account.

Only instances of same instance type & Availability Zone with Reserved Instances are valid for consolidated billing.

You can turn off Reserved Instance discount sharing on the Preferences page on the Billing and Cost Management console.

https://docs.aws.amazon.com/awsaccountbilling/latest/aboutv2/ri-behavior.html

https://docs.aws.amazon.com/awsaccountbilling/latest/aboutv2/consolidatedbilling-other.html

2 — You have implemented two AWS Direct Connect connections between your data center and AWS to have a stable and highly available network performance. After a recent IT financial audit, it was decided to review the current implementation and replace it with a more cost-effective option.

Which setup would you recommend for this scenario ?

A single AWS Direct Connect and an AWS managed VPN connection to connect your data center with Amazon VPC.

Note — Direct Connect does not provide any redundancy for its connection, it is recommended to set up at least two connections for high availability. However, this setup is expensive because the second connection is used only when the main connection fails which rarely happens.

To maintain high availability, but reduce the costs, you can use an AWS managed VPN connection as a backup, trading off with slower connection, until your Direct Connect connection is restored.

https://aws.amazon.com/premiumsupport/knowledge-center/configure-vpn-backup-dx/

3 — A data analytics company are in the business of collecting and processing vast amounts of data. Each data set generates up to thousands of files which can range from 10 MB to 1 GB in size. The archived data is rarely restored and in case there is a request to retrieve it, the company has a maximum of 24 hours to send the files. The data sets can be searched using its file ID, set name, authors, tags, and other criteria.

How can you provide the most cost-effective architecture to meet the above requirements ?

  • For each completed data set, compress and concatenate all of the files into a single Glacier archive.
  • Store the associated archive ID for the compressed files along with other search metadata in a DynamoDB table.
  • For retrieving the data, query the DynamoDB table for files that match the search criteria and then restore the files from the retrieved archive ID.

Note — You can further lower the cost of storing data by compressing it to a zip or tar file.

In addition, Glacier doesn’t have a built-in search function to help you retrieve the data. You have to store the archive ID to a database, such as DynamoDB, to help you effectively search the required data.

https://docs.aws.amazon.com/amazonglacier/latest/dev/glacier-select.html

4 — A startup is using multiple Amazon S3 buckets to store high-resolution media files. After deal with a leading media company, the two parties shall share those media files. The media company frequently performs multiple object retrievals from the S3 buckets every day, which increased the startup’s data transfer costs.

As the Solutions Architect, what can you do to help the startup lower their operational costs ?

Enable the Requester Pays feature of the startup’s S3 buckets to make the media company pay the cost of the data transfer from the buckets.

https://docs.aws.amazon.com/AmazonS3/latest/userguide/RequesterPaysBuckets.html

--

--